PCK1 Negatively Regulates Cell Cycle Progression and Hepatoma Cell

Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 https://doi.org/10.1186/s13046-019-1029-y

RESEARCH Open Access PCK1 negatively regulates cell cycle progression and hepatoma cell proliferation via the AMPK/p27Kip1 axis Lin Tuo†, Jin Xiang†, Xuanming Pan†, Jieli Hu, Hua Tang, Li Liang, Jie Xia, Yuan Hu, Wenlu Zhang, Ailong Huang*, Kai Wang* and Ni Tang*

Abstract Background: Altered glucose metabolism endows tumor cells with metabolic flexibility for biosynthesis requirements. Phosphoenolpyruvate carboxykinase 1 (PCK1), a key enzyme in the gluconeogenesis pathway, is downregulated in hepatocellular carcinoma (HCC) and predicts poor prognosis. Overexpression of PCK1 has been shown to suppress liver tumor growth, but the underlying mechanism remains unclear. Methods: mRNA and protein expression patterns of PCK1, AMPK, pAMPK, and the CDK/Rb/E2F pathway were determined using qRT-PCR and western blotting. Cell proliferation ability and cell cycle were assessed by MTS assay and flow cytometric analysis. The effect of PCK1 on tumor growth was examined in xenograft implantation models. Results: Both gain and loss-of-function experiments demonstrated that PCK1 deficiency promotes hepatoma cell proliferation through inactivation of AMPK, suppression of p27Kip1 expression, and stimulation of the CDK/Rb/E2F pathway, thereby accelerating cell cycle transitionfromtheG1toSphaseunderglucose-starved conditions. Overexpression of PCK1 reduced cellular ATP levels and enhanced AMPK phosphorylation and p27Kip1 expression but decreased Rb phosphorylation, leading to cell cycle arrest at G1. AMPK knockdown significantly reversed G1-phase arrest and growth inhibition of PCK1-expressing SK-Hep1 cells. In addition, the AMPK activator metformin remarkably suppressed the growth of PCK1-knockout PLC/PRF/5 cells and inhibited tumor growth in an orthotropic HCC mouse model. Conclusion: This study revealed that PCK1 negatively regulates cell cycle progression and hepatoma cell proliferation via the AMPK/p27Kip1 axis and supports a potential therapeutic and protective effect of metformin on HCC. Keywords: Phosphoenolpyruvate carboxykinase 1, Hepatocellular carcinoma, Cell cycle arrest, CDK/Rb/E2F pathway, AMP-activated protein kinase

Background of oxaloacetate (OAA) and GTP to phosphoenolpyruvate Metabolic reprogramming or “deregulating cellular ener- (PEP) and GDP by phosphoenolpyruvate carboxykinase getics” is a hallmark of cancer [1]. Tumor cells undergo (PCK; also known as PEPCK, EC number 4.1.1.32) is the aerobic glycolysis even in the presence of oxygen, also first rate-limiting step in hepatic gluconeogenesis [3, 4]. known as the Warburg effect, which is critical for meet- Two isoforms of PCK have been described, a cytoplas- ing the metabolic requirements of rapid cancer cell mic form (PCK1, PEPCK-C) and a mitochondrial iso- growth and proliferation [2]. The liver is the major site form (PCK2, PEPCK-M) [5]. In the mammalian liver, of gluconeogenesis during fasting, where the conversion PCK1 accounts for over 95% of gluconeogenesis activity [6], is upregulated in colon and melanoma cancers, and * Correspondence: [email protected]; [email protected]; is associated with increased glucose consumption to [email protected] † support anabolic metabolism and tumor growth. How- Lin Tuo, Jin Xiang and Xuanming Pan contributed equally to this work. ever, in hepatocellular carcinoma (HCC), the expression Key Laboratory of Molecular Biology for Infectious Diseases (Ministry of Education), Institute for Viral Hepatitis, Department of Infectious Diseases, levels of PCK and other key gluconeogenic enzymes are The Second Affiliated Hospital, Chongqing Medical University, Chongqing, strikingly suppressed [7, 8], indicating that PCK has People’s Republic of China

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 2 of 16

contrasting roles in tumorigenesis in different organs. Eagle’s medium (DMEM; Hyclone, Logan, UT, USA) Nevertheless, the role of PCK in HCC progression re- supplemented with 10% fetal bovine serum (FBS; Gibco, mains incompletely understood. Rockville, MD, USA), 100 IU penicillin, and 100 mg/mL Recent metabolomic studies confirmed that glucose levels streptomycin. HepG2 cells were cultured in minimum in tumors are generally lower than those in non-transformed essential medium (MEM; Hyclone). Cell lines involved tissues [9, 10]. However, cancer cells require sufficient glu- in this study were recently authenticated by short tan- cose to facilitate rapid tumor growth through the generation dem repeat (STR) profiling (Beijing Microread Gene of building blocks, which are necessary for the synthesis of Technology Co., Beijing, China). All cells were incubated essential cellular components [9, 11]. Therefore, cancer cells in a humidified atmosphere at 37 °C containing 5% CO2. require delicate energy sensing and metabolic adaptation For cell cycle analysis, cells were collected at different pathways. Mammalian AMP-activated protein kinase time points post-synchronization. For low-glucose treat- (AMPK) is a central cellular sensor of adenine nucleotides ment, cells were incubated in glucose-free DMEM that plays a key role in sensing cellular energy availability (Hyclone) supplemented with 10% FBS, 100 IU penicillin, and regulating metabolic adaptation pathways [11, 12]. En- 100 mg/mL streptomycin, and the indicated concentra- ergy stress conditions in tumor tissues are sensed by AMPK tions of glucose (Hyclone) for 12 h. For immunoblotting through elevated intracellularAMP/ATPratios.Onceacti- and cell cycle analysis under glucose-deprived conditions, vated, phosphorylated AMPK acts to restore the energy bal- cells were incubated in DMEM containing 1 mM glucose ance by increasing catabolic ATP-generating processes, such and metformin (5 mM; Cat# S1950; Selleck, Houston, TX, as glycolysis and fatty-acid oxidation, while decreasing USA) for 12 h and then collected for analysis. For prolifer- ATP-consuming processes, such as fatty acid synthesis and ation and colony formation assays, cells were cultured in cell proliferation [13]. AMPK activation has been reported to normal DMEM supplemented with 0.5 mM metformin. suppress cell proliferation by regulating cell cycle progression or inhibiting the synthesis of certain proteins, such as Plasmid constructs and adenovirus production p27Kip1 [14–16]. The full-length cDNA of PCK1 (coding sequence of In this study, we explored the potential role of PCK1 in NM_002591) was amplified from plasmid pOTB7-PCK1 the suppression of cellular proliferation and tumor growth (Cat# FL07339; GeneCopoeia, Rockville, MD, USA). in HCC. Interestingly, we found that PCK1 activates Primers are listed in Additional file 1: Table S1. The amp- AMPK and delays G1 to S phase (G1/S) transition. Our lified PCK1 fragment was inserted into the shuttle vector study also revealed that the AMPK activator metformin pAdTrack-TO4 (kindly provided by Dr. Tong-Chuan He, remarkably suppresses tumor growth in an orthotropic University of Chicago, USA). Adenoviral recombinant HCC mouse model. pAd-PCK1 was generated using the AdEasy system as described previously [17]. The analogous adenovirus Methods expressing GFP only (AdGFP) was used as control. Patient samples Liver tumor specimens and adjacent non-tumorous Lentivirus construction specimens were collected from 20 patients who under- Two pairs of oligonucleotides encoding short hairpin went surgical resection at the Second Affiliated Hospital RNA (shRNA) targeting isoform α1 of AMPK (Additional of Chongqing Medical University (Chongqing, China). file 1: Table S1) were designed and subcloned into the len- Patients had not received chemotherapy or radiation tiviral vector pLL3.7 (kindly gifted by Prof. Bing Sun from therapy before surgery. The samples were frozen imme- the Shanghai Institute of Biochemistry and Cell Biology, diately after surgery and stored in liquid nitrogen. All Chinese Academy of Sciences, China) to generate patients provided informed consent and the study was shAMPK lentivirus. A control virus (shControl), with a conducted with the approval of the Institutional Ethical scrambled shRNA insert, was also generated. For lentiviral Review Board of Chongqing Medical University (project supernatant production, HEK-293 T cells were transfected license number: 2017012). with 2 μgVSVG,3μg Δ8.9, and 4 μg of the shRNA lentiviral construct for lentiviral package using Lipofectamine Cell cultures and drug treatment 2000 reagent (Thermo Fisher Scientific, Waltham, MA, The human hepatoma cell lines HepG2, SK-Hep1, and USA), according to manufacturer’s instructions. SK-Hep1 PLC/PRF/5 were obtained from the American Type Cul- cells were transduced with the packaged lentiviruses in ture Collection (ATCC; Manassas, VA, USA). Other the presence of 5 μg/mL polybrene. hepatoma cell lines like Huh7 and normal human liver cells MiHA were obtained from the Cell Bank of the RNA-sequencing (RNA-seq) and expression analysis Chinese Academy of Sciences (Shanghai, China). Apart Hepatoma cells were infected with AdGFP or AdPCK1 for from HepG2, cells were cultured in Dulbecco’s modified 36 h and then RNA was extracted using TRIzol reagent Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 3 of 16

(Invitrogen, Rockville, MD, USA) according to manufac- GTX115461; GeneTex). Then, the membranes were turer’s instructions. RNA-seq and bioinformatic data incubated with the corresponding horseradish analysis were conducted at Shanghai Novel Bio Ltd. (Shang- peroxidase-conjugated secondary antibody (Abcam, Cam- hai, China). Strand-specific RNA-seq libraries were prepared bridge, UK). Proteins were detected using the Super Signal using the Total RNA-seq (H/M/R) Library Prep Kit (Vazyme West Pico Chemiluminescent Substrate Kit (Millipore) and Biotech, Nanjing, China) and were sequenced on an Ion quantified by densitometry using ImageJ software (National Torrent Proton Sequencer (Life Technologies, Carlsbad, Institutes of Health, Bethesda, MA, USA; http://imagej.nih.- CA, USA) according to the Ion PI Sequencing 200 Kit v2.0 gov/). GAPDH (Cat# AF0006; Beyotime) or β-actin (Cat# (Life Technologies). Raw reads in FASTQ format were BL005B; Biosharp) were used as internal control. All experi- quality-controlled using FastQC (http://www.bioinforma- ments were independently repeated three times. tics.babraham.ac.uk/projects/fastqc/). RNA-seq reads were aligned to the reference genome using Bowtie and uniquely Histological and immunohistochemistry (IHC) analysis mapped reads were used for further analysis. Gene expres- Liver samples were fixed in fresh 4% paraformaldehyde sion levels were expressed as reads per kilobase per million and subjected to routine histological procedures for em- reads (RPKM) and differences in gene expression were cal- bedding in paraffin. Then, the samples were cut in culatedwithrSeq(http://www-personal.umich.edu/~jian- 4.5 μm-thick sections that were processed for ghui/rseq/). hematoxylin and eosin (HE) or IHC staining with antibodies targeting PCK1 (1:500), pAMPK (1:200), or RNA extraction and real-time reverse transcription p27Kip1 (1:500). For IHC assays, the sections were incu- (qRT)-PCR bated with secondary anti-rabbit IgG (ZSGB-BIO, Total cellular RNA was extracted from cultured cells using Beijing, China) and stained with 3,3′-diaminobenzidine TRIzol reagent (Invitrogen) and then reverse transcribed (ZSGB-BIO). Stained slides were scanned with a Pan- using Moloney murine leukemia virus reverse transcriptase noramic Scan 250 Flash or MIDI system and images (Promega, Madison, WI, USA) and random hexamers were acquired using Pannoramic Viewer 1.15.2 (3DHis- (Promega). The synthesized cDNA was then used as tem- tech, Budapest, Hungary). plateforqPCRoftherespectivegenesusingSYBRGreen and a CFX Real-Time PCR Detection System (Bio-Rad La- boratories, Hercules, CA, USA). The actin beta gene CRISPR/Cas9-mediated knockout of PCK1 (ACTB) was used as a reference gene for normalization. The CRISPR/Cas9 plasmids lentiCRISPR v2, pMD2.G, –ΔΔ Relative mRNA levels were calculated using the 2 Ct and psPAX2 were kindly provided by Prof. Ding Xue from method. All primers are listed in Additional file 1:TableS1. the School of Life Sciences, Tsinghua University (Beijing, China). Single-guide RNAs targeting human PCK1 were Western blot analysis designed using the E-CRISP online tool (http://www.e-- Protein was extracted from cells or tissue samples using crisp.org/E-CRISP/designcrispr.html). The PCK1 targeting lysis buffer (Beyotime, Shanghai, China) supplemented with sequences (listed in Additional file 1: Table S1) were syn- 1 mM phenylmethylsulfonyl fluoride (Beyotime), after thesized by TsingKe Biological Technology (Chongqing, which the protein concentration was determined using a China) and cloned into lentiCRISPR v2 vectors. Lentivi- BCA protein assay kit (Beyotime). Proteins were resolved ruses were generated by co-transfecting HEK293T cells by sodium dodecyl sulfate polyacrylamide gel electrophor- with lentiCRISPR v2, envelop plasmid pMD2.G, and pack- esis and electrotransferred to polyvinylidene difluoride aging plasmid psPAX2 using Lipofectamine 2000 accord- membranes (Millipore, Billerica, MA, USA). The mem- ing to manufacturer’s instructions. After 48 h, PLC/PRF/5 branes were incubated with primary antibodies against cells were incubated with medium containing virus along PCK1 (1:1000; Cat# BS6870; Bioworld, Atlanta, GA, USA), with 5 μg/mL polybrene. Two days post-infection, cells phospho-AMPKα1 (T172; 1:1000; Cat# 2535; Cell Signaling were selected in the medium containing 2 μg/mL puro- Technology,Danvers,MA,USA),AMPKα (1:1000; Cat# mycin and then single-cell colonies were selected after 2795; Cell Signaling Technology), p27Kip1(1:1000; Cat# seeding in 96-well plates. For genotyping, clonal cell gen- 3686; Cell Signaling Technology), phospho-Rb (S780; omic DNA was extracted with a Genomic DNA Purifica- 1:1000; Cat# BS4164; Bioworld), Rb (T774; 1:1000; Cat# tion Kit (Genloci Biotechnologies Inc., Jiangsu, China) and BS1310; Bioworld), cyclin E1 (L389;1:1000; Cat# BS1086; then cloned into the pMD19-T TA cloning vector (Takara, Bioworld), CDK2 (78B2; 1:1000; Cat# 2546; Cell Signaling Kyoto, Japan) for sequencing. PCK1-knockout efficiency Technology), phospho-CDK2 (Thr160; 1:1000; Cat# 2561; was confirmed by western blotting. PCK1-knockout and Cell Signaling Technology), E2F2 (K236; 1:1000; Cat# control cells are hereafter referred to as PCK1-KO and BS2057; Bioworld), LKB1 (1:1000; Cat# GTX130697; parental cells, respectively. All primers are listed in Add- GeneTex, Irvine, CA, USA), and CAMKK2 (1:1000; Cat# itional file 1: Table S1. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 4 of 16

Proliferation assay day 7 after implantation, the mice were treated with Cells were seeded in 96-well plates at a density of 1 × metformin (250 mg/kg per day, intraperitoneally) or PBS 103 cells/well and cultured for 5 days. The absorbance at (equal volume, intraperitoneally) for 6 weeks. One 490 nm was measured in real time every 24 h after incu- mouse in a treatment group died due to postoperative bation with 20 μL of 3-(4,5-dimethylthiazol-2-yl)-5-(3-- infection during the experiment. Seven weeks after im- carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazo- plantation, the mice were sacrificed and liver tissues lium (MTS) labeling reagent solution (Promega) for 2 h. were collected for histological examination. All animal experiments were carried out according to Colony formation assay the guidelines of the Institutional Animal Care and Use Cells (1 × 103 cells/well) were seeded in 6-well plates and Committee at Chongqing Medical University (project cultured at 37 °C and 5% CO2 for 14 days. The medium license number: 2017012) and animal care and use proto- was replaced every three days. Cell colonies were washed cols adhered to national regulations for the administration twice with PBS, fixed with 4% paraformaldehyde for 30 of laboratory animals. min, and stained with 0.1% crystal violet for 20 min. The experiment was repeated at least three times. Statistical analysis Data are expressed as the mean ± standard deviation (SD). Cell cycle analysis ’ 6 Means were compared using Students t-test when com- Cells were harvested at a density of 1.0 × 10 cells/mL, paring two groups or one-way analysis of variance washed with PBS, and fixed with 70% ethanol at 4 °C (ANOVA) when comparing more than two groups. Corre- overnight. Then, the cells were washed with PBS, stained lations were assessed using the Spearman Rank Correl- with propidium iodide, and subjected to cell cycle ana- ation test. Two-sided P values < 0.05 were considered lysis using a FACSCalibur instrument (BD Biosciences, statistically significant. Statistical analyses were conducted Franklin Lakes, NJ, USA) and CellQuest software. using GraphPad Prism 7.0 software (La Jolla, CA, USA).

Determination of intracellular ATP levels Cellular ATP production was detected using an ATP Nucleotide sequence accession numbers Assay Kit (Beyotime) according to manufacturer’s instruc- The RNAseq data generated in this study have been sub- tions [18]. Briefly, cells were collected in a 1.5-mL tube mitted to the NCBI GEO database (http://www.ncbi.nlm.- and centrifuged. Then, 200 μL of lysis buffer from the nih.gov/geo) under the identifier GSE117822. ATP Assay Kit was added to each tube, after which the lysates were centrifuged at 12,000×g for 5 min. ATP produc- Results tion was measured by a luciferase assay of cell lysates and PCK1 suppresses hepatoma cell proliferation via G1/S normalized to cellular protein concentrations (nM ATP/ phase cell cycle arrest mg protein). Protein levels of the supernatant were mea- Previous studies have indicated that PCK1 expression is sured at 562 nm with a BCA assay kit (Beyotime). significantly lower in tumor tissues than in normal liver tissues and that downregulated PCK1 expression corre- Animal models lates with poor prognosis [18, 19]. To clarify the func- For the subcutaneous xenograft tumor model, 18 male tion of PCK1 during carcinogenesis and progression of BALB/c nude mice (5–6 weeks of age) were randomly di- HCC, we first examined endogenous PCK1 levels in sev- vided into three groups. MHCC-97H cells were eral hepatoma cell lines and MiHA cells. We found mock-infected or infected with AdGFP or AdPCK1 for 24 strong endogenous PCK1 expression in PLC/PRF/5, h, then collected for subcutaneous injection (1 × 105 cells/ HepG2, and MiHA, modest expression in Huh7, and injection) into the flanks of athymic BALB/c nude mice. low expression levels in SK-Hep1 and MHCC-97H cells Tumor volume was monitored by measuring the length (Additional file 2: Figure S1a). According to the en- (L) and width (W) at 3-day intervals for 5 weeks. Tumor dogenous PCK1 expression pattern in hepatoma cells, volume [cm3] was calculated as L [cm] × (square of W we evaluated the effect of PCK1-overexpression (OE) in [cm2])/2. After 5 weeks, the mice were sacrificed and SK-Hep1, Huh7, and MHCC-97H cells as well as the ef- tumor tissues were collected for histological analysis. fect of PCK1-knockout (KO) in PLC/PRF/5 cells. Over- For the orthotopic implantation model, 15 BALB/c expression of PCK1 significantly suppressed proliferation nude mice were randomly divided into parental, of SK-Hep1, MHCC-97H, and Huh7 cells compared PCK1-KO, and metformin-treated PCK1-KO groups with that of GFP-expressing control cells (Fig. 1a, b, and (five mice per group). The PLC/PRF/5 parental and Additional file 2: Figure S1b, S1c). On the other hand, PCK1-KO cells (1 × 105 cells/injection) were collected PCK1 knockout promoted cell proliferation in PLC/ and implanted into the left lobes of nude mice livers. On PRF/5 cells (Fig. 1c and d). Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 5 of 16

Fig. 1 (See legend on next page.) Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 6 of 16

(See figure on previous page.) Fig. 1 Suppression of PCK1 expression promotes cell proliferation by inducing G1/S phase transition in hepatoma cells. a-d Cell proliferation curves and colony formation assay. a, b SK-Hep1 and MHCC-97H cells were infected with adenoviruses expressing PCK1 or GFP control. PCK1 overexpression (PCK1-OE) was confirmed by western blotting. c, d Endogenous PCK1 was knocked out in PRF/PLC/5 cells by CRISPR-Cas9 and confirmed by western blotting. Cells were seeded into 96-well plates at 1 × 103 cells/mL and counted every 24 h in triplicate for cell proliferation analysis. Cells were seeded in 6-well plates at 1 × 103 cells/well and cultured for 2 weeks for colony formation assays. e, f Flow cytometric analysis. e PCK1-OE SK-Hep1 cells and f PCK1-KO cells were collected at different time points and subjected to flow cytometry. Populations of cells at G1 and S phases are indicated as percentages of the whole cell population. Data represent the mean ± SD of three independent experiments; (means ± SD; *P < 0.05, **P < 0.01)

To determine whether the impact of PCK1 on cell pro- (Fig. 2e). Together, these findings suggest that PCK1 may liferation is associated with cell cycle control, we per- dampen CDK/Rb/E2F signaling to inhibit cell cycle pro- formed flow cytometric analyses of PCK1-overexpression gression, resulting in the suppression of cell proliferation. (OE) or PCK1-knockout (KO) hepatoma cells. Cells were cultured in normal DMEM and collected continuously PCK1 arrests the cell cycle through AMPK activation during days 1–5 post-synchronization. Time-course ana- PCK1 catalyzes the formation of PEP from OAA, this reac- lysis indicated that PCK1 overexpression significantly tion—along with GTP consumption—may result in cellular inhibited G1/S transition in SK-Hep1 cells after prolonged energy reduction [20]. Therefore, we evaluated whether incubation, especially 3–5 days after synchronization treat- AMPK, an important cellular energy sensor, is involved in ment (Fig. 1e and Additional file 3: Figure S2a), whereas cell cycle regulation by PCK1. We first monitored the pro- PCK1 deficiency promoted G1/S transition in PLC/PRF/5 tein levels of AMPK and phospho-AMPK (Thr 172; cells 3–5 days post-synchronization (Fig. 1f and Additional pAMPK) in PCK1-OE and PCK1-KO hepatoma cells under file 3: Figure S2b). Taken together, these results indicate normal culture conditions at different time points. pAMPK that PCK1 inhibits hepatoma cell proliferation by delaying levels were significantly upregulated in PCK1-OE cells after G1/S transition. 5-day culture, with higher levels of pAMPK detected after 7-day culture. (Fig. 3a). Conversely, knockout of PCK1 in PCK1 inhibits the CDK/Rb/E2F signaling pathway PRF/PLC/5 cells inhibited pAMPK expression after 5 or To explore the mechanism underlying PCK1-induced 7-day cultures compared with that of parental cells (Fig. G1 cell cycle arrest, we performed genome-wide 3b). Considering that nutrients such as glucose in the cul- RNA-Seq of Huh7 cells infected with AdPCK1 or ture medium are gradually consumed during long-term AdGFP control (Fig. 2). Analysis of the differentially culture, we compared AMPK and pAMPK expression expressed genes (DEGs) indicated that multiple inhibi- levels in PCK1-OE and PCK1-KO hepatoma cells cultured tors of cell cycle progression, especially G1-S phase in- under low-glucose (1 mM) or normal-glucose (25 mM) hibitors including CDKN2B (p15), CDKN1B (p27), conditions for 12 h. Consistent with the AMPK expression CDK4, and CDK9, were strongly upregulated by PCK1. profile in long-term cell culture, PCK1 overexpression Meanwhile, PCK1 dramatically downregulated the cell markedly upregulated pAMPK expression (Fig. 3c), whereas cycle accelerators E2F2, CCNE1, and CCND1 (Fig. 2a). knockout of PCK1 downregulated pAMPK expression (Fig. These RNA-seq results were validated by qRT-PCR. 3d) in cells under low-glucose conditions. mRNA levels of CDK4, MCM2, PCNA, CCNE1, and AMPK is mainly activated via phosphorylation through E2F2 were found downregulated, whereas those of the increased expression of liver kinase B1 (LKB1), the CDKN1B (p27Kip1) were enhanced by PCK1. In contrast, Ca2+/calmodulin-dependent protein kinase kinase2 PCK1 knockout resulted in reversed regulatory effects (CaMKK2), and an elevation in cellular AMP/ATP ratios on these molecules (Fig. 2b and c). [21]. Therefore, we further explored the mechanism re- Next, we conducted immunoblot analysis to determine sponsible for AMPK activation by PCK1. Immunoblot as- the impact of PCK1 on the protein levels of cell says showed that protein levels of LKB1 and CaMKK2 cycle-related proteins, especially G1/S cell cycle were not affected upon modulation of PCK1 expression checkpoint-related proteins. We found that the p27Kip1 (Fig. 3e and f), indicating that LKB1 and CaMKK2 are not protein was dramatically upregulated in AdPCK1-infected largely responsible for PCK1-mediated AMPK activation. SK-Hep1 cells, whereas pRB, pCDK2, and cyclin E1 were However, ATP levels in PCK1-OE SK-Hep1 cells were sig- markedly downregulated in PCK1-OE cells (Fig. 2d). In nificantly reduced after 5 and 7-day culture (Fig. 3g) and contrast, PCK1 knockout increased the protein levels of markedly increased in PCK1-KO PLC/PRF/5 cells after 3, pRB, pCDK2, E2F2, and cyclin E1 in PLC/PRF/5 cells, 5, and 7-day culture (Fig. 3h), which is highly consistent whereas p27Kip1 expression was significantly decreased with the time-course analysis of AMPK activation. These Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 7 of 16

Fig. 2 (See legend on next page.) Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 8 of 16

(See figure on previous page.) Fig. 2 PCK1 controls the expression of cell cycle-related genes. a Heat map representing the differential expression of E2F family members, transcription factors, and E2F target genes as determined by RNA-Seq in PCK1-OE and GFP control cells. Hepatoma cells infected with adenoviruses expressing PCK1 (AdPCK1) or vector control (AdGFP) were subjected to RNA-Seq analysis. b–e Differentially expressed genes (DEGs) were validated in PCK1-OE and PCK1-KO hepatoma cells by RT-qPCR and western blot analysis. PCK1-OE SK-Hep1 cells and PCK1-KO cells were treated as described in Fig. 1. b, c mRNA expression levels of the indicated genes were measured by qRT-PCR. d, e Western blot analysis of the indicated genes in PCK1-OE and PCK1-KO cells. β-actin was used as a loading control. Relative levels of target proteins were measured by densitometry. Data are presented as the mean ± SD of three independent experiments; *P < 0.05, **P < 0.01

results indicate that a decrease in ATP generation plays an Metformin-mediated AMPK activation suppresses liver essential role in PCK1-induced AMPK activation. tumor growth in nude mice Considering that the murine subcutaneous xenograft model using PLC/PRF/5 cell is difficult to achieve, we ex- PCK1 delays G1/S phase transition by upregulating amined the effects of PCK1 deficiency on tumor growth p27Kip1 expression via AMPK activation in an orthotopic transplantation HCC mouse model. Considering that AMPK activation is involved in PCK1-KO and parental PLC/PRF/5 cells were trans- PCK1-induced growth inhibition, we explored whether planted into the livers of nude mice. Compared with par- knockdown of AMPK can partially reverse the cell cycle ental PLC/PRF/5 cells, PCK1-KO significantly promoted arrest and growth-retarding effects of PCK1. We sup- orthotopic tumor formation in mice (Fig. 6a–c). pressed AMPK expression by infecting SK-Hep1 cells To investigate whether metformin reverses the PCK1 with lentiviruses expressing shRNAs targeting AMPK deficiency-induced tumor-promoting effects, mice ortho- (shAMPK1# and shAMPK2#; Additional file 4: Figure topically implanted with PCK1-KO hepatoma cells were S3a). As expected, AMPK silencing strongly decreased divided into two groups, one group was treated with met- p27Kip1 expression and increased pRb expression under formin (250 mg/kg/day) and the other received PBS as low-glucose conditions (1 mM glucose; Additional file 4: control (Fig. 6a). We found that metformin treatment dra- Figure S3b). In addition, knockdown of AMPK pre- matically repressed tumor growth in mice (Fig. 6bandc). vented PCK1-induced upregulation of p27Kip1, downreg- HE staining revealed that livers of the PCK1-KO group ulation of pRb, delay of G1/S transition, and inhibition displayed many pleomorphic and atypical hepatocytes as of cell proliferation (Fig. 4, and Additional file 5: Figure well as remarkably altered nodular liver structure, which S4a). The AMPK activator metformin partially offset the were attenuated by metformin treatment (Fig. 6d). Fur- PCK1 deficiency-mediated downregulation of p27Kip1, thermore, pAMPK expression was significantly higher in upregulation of pRb, promotion of G1/S transition, and the metformin-treated group than in the control group acceleration of hepatoma cell proliferation (Fig. 4, and (Fig. 6e), and this result was consistent with our in vitro Additional file 5: Figure S4b). These results indicate that findings. Taken together, these results demonstrate that PCK1 induces p27Kip1 expression through activation of metformin treatment efficiently inhibits the tumorigenesis the AMPK pathway, resulting in suppression of the G1/ of PCK1-KO hepatoma cells in nude mice through the S transition and eventual inhibition of hepatoma cell AMPK/p27Kip1 axis. proliferation. Downregulation of PCK1 is positively correlated with PCK1 acts as a tumor suppressor of HCC in vivo pAMPK and p27Kip1 expression in HCC patients Next, we evaluated the effect of PCK1 gain-of-function Finally, we analyzed PCK1 expression in clinical HCC on hepatoma cell growth in a murine subcutaneous samples and explored the potential association between xenograft model. As shown in Fig. 5a, b, and Additional PCK1 and p27Kip1 expression in HCC tissues. In an in- file 6: Figure S5a, tumor growth was significantly sup- dependent cohort of 371 HCC samples from the Cancer pressed in nude mice that had received PCK1-OE cells. Genome Atlas (TCGA) database, PCK1 expression was Furthermore, tumor volume and weight were notably re- positively correlated with p27Kip1 expression (r = 0.3158, duced in the PCK1-OE group compared with the GFP P < 0.0001; Fig. 7a) and negatively correlated with control group (Fig. 5c and d). Moreover, pAMPK and CCNE1 expression (r = 0.3692, P < 0.0001; Fig. 7b). In p27Kip1 expression was markedly increased in tumor tis- addition, we examined the protein levels of PCK1, sues exogenously expressing PCK1, as evidenced by pAMPK, and p27Kip1 in paired HCC and adjacent western blotting and IHC (Fig. 5e and Additional file 6: non-cancerous tissues from 20 patients. PCK1 protein Figure S5b). Collectively, these findings indicate that levels were markedly lower in 16 of the 20 HCC tissues PCK1 inhibits tumor growth both in vitro and in vivo. (85%; Fig. 7c) compared with non-cancerous tissues. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 9 of 16

Fig. 3 (See legend on next page.) Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 10 of 16

(See figure on previous page.) Fig. 3 PCK1 activates AMPK by decreasing ATP levels in glucose-deprived hepatoma cells. a–d Immunoblot analysis of AMPK and phosphorylated AMPK. (a) PCK1-OE or (b) PCK1-KO hepatoma cells were cultured in normal DMEM for 1, 3, 5, or 7 days, while (c) PCK1-OE or (d)PCK1-KOhepatoma cells were cultured in DMEM containing different concentrations of glucose (25 or 1 mM) for 12 h. AMPK phosphorylation was measured by western blot analysis. Results are representative of three independent experiments. e, f Protein expression of LKB1 and CaMKK2 in (e) PCK1-OE and (f) PCK1-KO hepatoma cells. β-actin was used as a loading control. Integrated density of pAMPK/AMPK, LKB1, and CaMKK2 was quantitatively analyzed using ImageJ software and the results were normalized to mock or parental group. *P <0.05,**P < 0.01. (g, h) Quantification of ATP concentrations in (g) PCK1-OE and (h) PCK1-KO HCC cells cultured in normal DMEM medium for 1, 3, 5, or 7 days. Cellular ATP levels were determined as described in the Methods section. Values represent mean ATP/protein levels ± SD of triplicate samples. *P < 0.05, **P <0.01

Furthermore, pAMPK and p27Kip1 levels were also re- signaling pathway. Mechanistically, PCK1 decreases cel- duced in 14 and 13 of the 16 PCK1-downregulated HCC lular ATP levels and activates the AMPK/p27Kip1 axis in tissues, respectively. (Fig. 7c). Downregulation of PCK1 glucose-deprived hepatoma cells. Interestingly, metfor- and pAMPK expression in HCC tissues was confirmed min, an AMPK activator, restrains the tumorigenesis of by IHC staining (Fig. 7d). These results reveal that PCK1-deficient HCC. Based on our findings, we suggest downregulation of PCK1 is linked to decreased levels of a potential model for AMPK activation and cell cycle ar- pAMPK and p27Kip1 in HCC tissues, suggesting that rest in the PCK1-induced inhibition of hepatoma cell PCK1 deficiency may be involved in HCC progression proliferation (Fig. 7e). through inhibition of the AMPK pathway and repression AMPK is recognized as a cellular energy sensor that is of p27Kip1 expression. activated by phosphorylation at Thr172 [30]. Once activated, AMPK triggers catabolic pathways to generate ATP, Discussion while inhibiting anabolic ATP-consuming pathways in an Hepatic gluconeogenesis is an enzymatic process that attempt to restore cellular energy homeostasis [13, 31, 32]. ensures glucose production rates exactly match AMPK is known to inhibit gluconeogenesis by phosphor- whole-body glucose requirements; PCK1 is regarded as ylating CREB-regulated transcription coactivator 2 [33] one of the key enzymes involved in this process [22]. Re- and class IIA histone deacetylases [34]—co-activators of cent studies have shown that PCK1 functions as a regu- the CREB and FOXO pathways, respectively [35]. Interest- lator of hepatic energy metabolism and gluconeogenesis, ingly, a recent study reported that the glycolytic enzymes with its dysregulation linked to metabolic diseases, such aldolase and fructose-1,6-bisphosphate can suppress as diabetes, obesity, insulin resistance, and tumor devel- AMPK activation [36]. The present study demonstrated opment [23–25]. However, PCK1 has contrasting effects that the gluconeogenic enzyme PCK1 also plays an im- on tumorigenesis in different tumor types. In colorectal portant role in AMPK activation. As previously reported, cancer and melanoma, increased PCK1 expression is as- AMP binding to AMPK γ-subunits causes AMPK allo- sociated with upregulated anabolic metabolism and cell steric activation [37, 38], while upstream kinases such as proliferation, indicating that PCK1 functions as an onco- LKB1 [39]andCaMKK2[40]activateAMPKthrough gene in these cancers [20, 25, 26]. In contrast, Zeribe et modulating the phosphorylation state of AMPK α-subunit al. [27] observed that PCK1 ranked in the top 25 down- at Thr172 [21]. Our study showed that upon modulation regulated genes in at least six different HCC datasets. of PCK1 expression, protein levels of LKB1 and CaMKK2 Similarly, PCK1 downregulation has been observed in were not affected while ATP levels were significantly al- other cohorts of HCC, where restoration of its expres- tered. On the other hand, PCK1 was reported to consume sion leads to tumor suppression [18, 19, 28], indicating GTP when catalyzing the formation of PEP from OAA that PCK1 acts as a tumor suppressor in HCC. However, [41]. Taking into consideration the conversion of GTP the mechanisms underlying the crosstalk of PCK1 with to ATP via the nucleoside diphosphate kinase reaction oncogenic pathways in the modulation of cellular behav- [42, 43], PCK1 may be involved in the regulation of ior, such as cell apoptosis and proliferation, remain intracellular ATP levels. In support of our findings, poorly understood. Liu et al. [18] recently reported that PCK1 activated The present study demonstrated that PCK1 deficiency AMPK upon glucose deprivation in HCC. Their study promotes HCC growth both in vitro and in vivo, revealed that forced PCK1 expression induced energy whereas overexpression of PCK1 represses HCC growth, crisis and oxidative stress, leading to cell apoptosis which is consistent with previous findings [18, 29]. Fur- under low-glucose conditions; however, the direct thermore, transcriptome and cell cycle analysis revealed consequences of AMPK activation were not illumi- that PCK1 may suppress G1/S transition and cell prolif- nated in depth. Here, we demonstrated that eration in hepatoma cells via inhibiting the CDK/Rb/E2F PCK1-induced AMPK activation increases p27Kip1 Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 11 of 16

Fig. 4 PCK1-mediated AMPK activation and delayed cell proliferation depend on p27Kip1 expression and phosphorylation of Rb protein. a, b Protein levels of the indicated proteins in PCK1-OE and KO hepatoma cells. β-actin was used as a loading control. a AMPK-stable knockdown SK- Hep1 cells (SK-Hep1/shAMPK) or control cells (SK-Hep1/shControl) were infected with AdPCK1 or AdGFP for 48 h, respectively. b PCK1-KO or parental PLC/PRF/5 cells were treated with 5 mM metformin for 12 h. Cell lysates were then subjected to immunoblot assays with the indicated antibodies for detecting the protein expression levels of p27Kip1, pRB, AMPK, and pAMPK. Results shown are representative samples from at least three independent experiments. Integrated density of these proteins was quantitatively analyzed and the results normalized to the shCon+AdGFP or parental group. c, d Flow cytometric analysis. Cells were treated as described above, collected at different timepoints (1, 2, 3, 4, or 5 days), and subjected to flow cytometry. Percentages of cells in the G1 and S phases are shown within each graph. Data represent the mean ± SD of three independent experiments. *P < 0.05. e–h Cell proliferation curves and colony formation assay. SK-Hep1 cells infected with shAMPK or shControl lentivirus were treated as described in a while PCK1-KO and parental PLC/PRF/5 cells were treated with 0.5 mM metformin. e, f Representative images and quantification of colony formation. Cell colonies were counted after 2 weeks of incubation. g, h Cell proliferation curves. Cells were counted every 24 h in triplicate. Values represent the mean ± SD of three independent experiments. *P < 0.05, **P < 0.01 Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 12 of 16

Fig. 5 PCK1 inhibits hepatoma growth in a subcutaneous xenograft mouse model. MHCC-97H cells were infected with AdGFP or AdPCK1 for 24 h. Control MHCC-97H cells and infected cells (1 × 105 cells/injection) were injected into the flank of nude mice (n = 6 per group). a Tumor growth curves of subcutaneous xenograft tumor model mice. Tumor volumes were calculated as M × (square of W)/2. *P < 0.05 by two-tailed unpaired Student’s t-test. Mice were sacrificed 38 days after injection, followed by xenograft tumor dissection. b Representative images of xenografts tumors. c Tumor size and d tumor weight of subcutaneous xenografts. e PCK1, pAMPK, and p27Kip1 protein expression in tumor tissue samples as detected by western blotting. Relative levels of the indicated proteins were quantified using ImageJ software. **P < 0.01

expression and suppresses hepatoma cell proliferation. anti-tumor drug. Metformin reportedly inhibits tumor Several studies have suggested that AMPK arrests cell development, progression, and metastasis by reducing cycle progression through the upregulation of G1/S cell proliferation [46], reversing epithelial-mesenchymal phase transition inhibitors such as p27Kip1 [16, 44, transition [47], and enhancing cell sensitivity to chemo- 45]. Activation of AMPK may upregulate p27Kip1 ex- therapeutic drugs [48]. The anti-tumor effect of met- pression directly or through the AMPK-p53-p27 axis. formin has been verified in breast, lung, and gastric [15, 16] Moreover, p27 Kip1 stability via phosphoryl- cancers [46, 49, 50]. A recent study indicated that ation is also increased by activation of AMPK signal- type 2 diabetes is correlated with an increased risk ing [44]. of HCC and that metformin decreases HCC risk in Metformin, a lipophilic biguanide, is used in first-line diabetic patients in a dose-dependent manner [51]. pharmacotherapy for type 2 diabetes. Increasing evi- Our study corroborated the anti-tumor effect of met- dence suggests that it may also function as an formin on HCC from the metabolic perspective. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 13 of 16

Fig. 6 Metformin suppresses tumor growth in an orthotopic implantation tumor model established with PCK1-KO hepatoma cells. PCK1-KO and parental cells (1 × 105 cells/injection) were implanted into the left lobes of nude mice livers (n = 5 per group). One week after implantation, mice were treated with metformin (250 mg/kg per day, intraperitoneally) or PBS (equal volume, intraperitoneally) for 6 weeks. The mice were sacrificed 7 weeks after implantation. a Schematic overview of the experiment design. b Tumor foci number in the liver of each mouse. c Representative gross appearance of murine livers. Red arrows: metastatic foci. d, e Representative d HE staining and e IHC detection of PCK1 and pAMPK in liver tumor tissues. Magnification: 100×, 200×

Conclusions vivo, highlighting its potential therapeutic and protective The present study revealed that the gluconeogenic en- effects on HCC. Our findings provide new insights into zyme PCK1 negatively regulates cell cycle progression the mechanism of uncontrolled cell proliferation medi- and cellular proliferation via the AMPK/p27Kip1 axis, in- ated by altered metabolism in liver cancer. Future stud- dicating a tumor suppressor role for PCK1 in HCC. Not- ies on PCK1 dysregulation and hepatocarcinogenesis will ably, the AMPK activator metformin restrained the provide novel diagnostic and therapeutic strategies tumorigenesis of PCK1-deficient HCC in vitro and in against HCC. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 14 of 16

Fig. 7 PCK1 expression is reduced in HCC tissues and is positively correlated with pAMPK and p27Kip1 expression. a, b Correlation between CDKN1B, CCNE1, and PCK1. Correlation between mRNA expression levels of a CDKN1B or b CCNE1 and PCK1 based on data from 371 HCC patients from the TCGA database mRNA dataset. Statistical analysis was performed using Pearson’s correlation coefficient (P < 0.0001). c PCK1, pAMPK, and p27Kip1 protein levels in 20 paired primary HCC tissues and adjacent non-tumor tissues. β-actin was used as a loading control. d Representative IHC images of PCK1 and pAMPK in HCC and adjacent non-tumor tissues. Magnification, 200×. e Proposed model for activation of AMPK and cell cycle arrest in PCK1-induced inhibition of hepatoma cell growth. The gluconeogenic enzyme PCK1 catalyzes an energy-consuming reaction, which consumes GTP to generate GDP. PCK1 induces AMPK activation via ATP-dependent mechanisms both in hepatoma cells and in vivo. AMPK activation further promotes G1/S transition by regulating p27Kip1 expression and negatively modulating pCDK2, cyclin E1, and pRB expression. Eventually, PCK1 inhibits the growth of hepatoma cells and the development of HCC through AMPK activation and cell cycle arrest Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 15 of 16

Additional files Received: 24 October 2018 Accepted: 7 January 2019

Additional file 1: Primer sequences of target genes. (XLSX 13 kb)

Additional file 2: PCK1 supresses Huh7 cells proliferation. (JPG 2932 kb) References Additional file 3: Knockout of PCK1 promotes G1/S phase transition in 1. Hanahan D, Weinberg RA. Hallmarks of Cancer: the next generation. Cell. hepatoma cells. (JPG 4192 kb) 2011;144:646–74. Additional file 4: AMPK regulates the expression of p27 and pRB in 2. Hay N. Reprogramming glucose metabolism in cancer: can it be exploited glucose-deprived hepatoma cells. (JPG 2140 kb) for cancer therapy? Nat Rev Cancer. 2016;16:635–49. 3. Hanson RW, Garber AJ. Phosphoenolpyruvate carboxykinase. I. Its role in Additional file 5: PCK1 retards G1/S phase transition via AMPK gluconeogenesis. Am J Clin Nutr. 1972;25:1010–21. phosphorylation. (JPG 4500 kb) 4. Rognstad R. Rate-limiting steps in metabolic pathways. J Biol Chem. 1979; Additional file 6: PCK1 inhibits hepatoma growth in a subcutaneous 254:1875–8. xenograft mouse model. (JPG 3699 kb) 5. Nordlie RC, Lardy HA. Mammalian liver Phosphoenolpyruvate Carboxykinase activities. J Biol Chem. 1963;238:2259–63. 6. Méndez-Lucas A, Duarte JAG, Sunny NE, Satapati S, He T, Fu X, et al. PEPCK- Abbreviations M expression in mouse liver potentiates, not replaces, PEPCK-C mediated AMPK: 5’-AMP-activated protein kinase; ATP: Adenosine triphosphate; gluconeogenesis. J Hepatol. 2013;59:105–13. CDKs: Cyclin-dependent kinases; HCC: Hepatocellular carcinoma; 7. Ma R, Zhang W, Tang K, Zhang H, Zhang Y, Li D, et al. Switch of glycolysis PCK1: Phosphoenolpyruvate carboxykinase 1; Rb: Retinoblastoma protein to gluconeogenesis by dexamethasone for treatment of hepatocarcinoma. Nat Commun. 2013 [cited 2018 Jan 7];4. Available from: http://www.nature. Acknowledgements com/doifinder/10.1038/ncomms3508 We would like to thank Dr. T.-C He (University of Chicago, USA) for providing 8. Hirayama A, Kami K, Sugimoto M, Sugawara M, Toki N, Onozuka H, et al. the plasmids pAdEasy system. We are also grateful to Prof. Ding Xue (School Quantitative metabolome profiling of colon and stomach cancer of Life Sciences, Tsinghua University) for supplying the CRISPR/Cas9 system, microenvironment by capillary electrophoresis time-of-flight mass – lentiCRISPR v2, pMD2.G and psPAX2. spectrometry. Cancer Res. 2009;69:4918 25. 9. Birsoy K, Possemato R, Lorbeer FK, Bayraktar EC, Thiru P, Yucel B, et al. Metabolic determinants of cancer cell sensitivity to glucose limitation and Funding biguanides. Nature. 2014;508:108–12. This work was supported by research grants from China National Natural 10. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg Science Foundation (grant nos. 81602417 to KW, 81872270, 81572683, effect: the metabolic requirements of cell proliferation. Science. 2009;324: 81371827, and 31171307 to NT, 81471946, 81661148057 to AH, and 1029–33. – 81671997 to JH), the Major National S&T program (2017ZX10202203 004, 11. Lin S-C, Hardie DG. AMPK: sensing glucose as well as cellular energy status. – 2012ZX10002005 003-002 to NT), Natural Science Foundation Project of CQ Cell Metab. 2018;27:299–313. CSTC (cstc2018jcyjAX0254 to N.T.), The Program for Innovation Team of 12. Endo H, Owada S, Inagaki Y, Shida Y, Tatemichi M. Glucose starvation Higher Education in Chongqing (grant no. CXTDX201601015), the Leading induces LKB1-AMPK-mediated MMP-9 expression in cancer cells. Sci Rep. Talent Program of CQ CSTC (grant no. CSTCCXLJRC201719 to NT), the 2018 [cited 2018 Jul 22];8. Available from: http://www.nature.com/articles/ Scientific Research Innovation Project for Postgraduate in Chongqing (grant s41598-018-28074-w no. YB17116), and Talent Development Program of CQMU for Postgraduate 13. Hardie DG, Ross FA, Hawley SA. AMPK: a nutrient and energy sensor that (grant no. BJRC201705). maintains energy homeostasis. Nat Rev Mol Cell Biol. 2012;13:251–62. 14. Motoshima H, Goldstein BJ, Igata M, Araki E. AMPK and cell proliferation - Availability of data and materials AMPK as a therapeutic target for atherosclerosis and cancer: AMPK inhibits All data used in this study are fully available without restrictions. cell proliferation. J Physiol. 2006;574:63–71. 15. Wang S-T, Ho HJ, Lin J-T, Shieh J-J, Wu C-Y. Simvastatin-induced cell cycle arrest through inhibition of STAT3/SKP2 axis and activation of AMPK to Authors’ contributions promote p27 and p21 accumulation in hepatocellular carcinoma cells. Cell NT, KW and AH conceived and designed the study; LT, JX, XP, YH and LL Death Dis. 2017;8:e2626. performed experimental work; JH, HT, WZ and JX performed data analysis; LT 16. Qi H, Liu Y, Li S, Chen Y, Li L, Cao Y, et al. Activation of AMPK attenuated prepared the manuscript; KW and NT provided administrative support and cardiac fibrosis by inhibiting CDK2 via p21/p27 and miR-29 family pathways funded experiments; all authors read and approved the final manuscript. in rats. Mol Ther - Nucleic Acids. 2017;8:277–90. 17. Gao Q, Wang K, Chen K, Liang L, Zheng Y, Zhang Y, et al. HBx protein- Ethics approval and consent to participate mediated ATOH1 downregulation suppresses ARID2 expression and All patients were provided informed consent, and this study was conducted promotes hepatocellular carcinoma. Cancer Sci. 2017;108:1328–37. with the approval of Institutional Ethical Review Board of Chongqing Medical 18. Liu M-X, Jin L, Sun S-J, Liu P, Feng X, Cheng Z-L, et al. Metabolic University (project license number: 2017012). All animal experiments were reprogramming by PCK1 promotes TCA cataplerosis, oxidative stress and conducted according to the guidelines of the Institutional Animal Care and apoptosis in liver cancer cells and suppresses hepatocellular carcinoma. Use Committee at Chongqing Medical University (The project license Oncogene. 2018;37:1637–53. number: 2017012), and protocols of animal care and use adhere to National 19. Tang Y, Zhang Y, Wang C, Sun Z, Li L, Cheng S, et al. Overexpression of Regulations of the administration of laboratory animals. PCK1 gene antagonizes hepatocellular carcinoma through the activation of gluconeogenesis and suppression of glycolysis pathways. Cell Physiol Biochem. 2018:344–55. Consent for publication All contributing authors agree to the publication of this article. 20. Park J-W, Kim SC, Kim WK, Hong JP, Kim K-H, Yeo HY, et al. Expression of phosphoenolpyruvate carboxykinase linked to chemoradiation susceptibility of human colon cancer cells. BMC Cancer. 2014 [cited 2018 May 13];14. Available Competing interests from: http://bmccancer.biomedcentral.com/articles/10.1186/1471-2407-14-160 The authors declare that they have no competing of interest. 21. Ha J, Guan K-L, Kim J. AMPK and autophagy in glucose/glycogen metabolism. Mol Asp Med. 2015;46:46–62. 22. Vieira P, Cameron J, Rahikkala E, Keski-Filppula R, Zhang L-H, Santra S, et al. Novel Publisher’sNote homozygous PCK1 mutation causing cytosolic phosphoenolpyruvate carboxykinase Springer Nature remains neutral with regard to jurisdictional claims in deficiency presenting as childhood hypoglycemia, an abnormal pattern of urine published maps and institutional affiliations. metabolites and liver dysfunction. Mol Genet Metab. 2017;120:337–41. Tuo et al. Journal of Experimental & Clinical Cancer Research (2019) 38:50 Page 16 of 16

23. Beale EG, Harvey BJ, Forest C. PCK1 and PCK2 as candidate diabetes and 46. Han G, Gong H, Wang Y, Guo S, Liu K. AMPK/mTOR-mediated inhibition of obesity genes. Cell Biochem Biophys. 2007;48:89–95. survivin partly contributes to metformin-induced apoptosis in human 24. Satapati S, Kucejova B, Duarte JAG, Fletcher JA, Reynolds L, Sunny NE, et al. gastric cancer cell. Cancer Biol Ther. 2015;16:77–87. Mitochondrial metabolism mediates oxidative stress and inflammation in 47. Qu C, Zhang W, Zheng G, Zhang Z, Yin J, He Z. Metformin reverses fatty liver. J Clin Invest. 2016;126:1605. multidrug resistance and epithelial-mesenchymal transition (EMT) via 25. Li Y, Luo S, Ma R, Liu J, Xu P, Zhang H, et al. Upregulation of cytosolic activating AMP-activated protein kinase (AMPK) in human breast cancer phosphoenolpyruvate carboxykinase is a critical metabolic event in cells. Mol Cell Biochem. 2014;386:63–71. melanoma cells that repopulate tumors. Cancer Res. 2015;75:1191–6. 48. Uehara T, Mitsuhashi A, Tsuruoka N, Shozu M. Metformin potentiates the 26. Montal ED, Dewi R, Bhalla K, Ou L, Hwang BJ, Ropell AE, et al. PEPCK anticancer effects of cisplatin under normoxic conditions in vitro. Oncol coordinates the regulation of central carbon metabolism to promote Rep. 2015;33:744–50. Cancer cell growth. Mol Cell. 2015;60:571–83. 49. Li L, Han R, Xiao H, Lin C, Wang Y, Liu H, et al. Metformin sensitizes EGFR- 27. Nwosu ZC, Megger DA, Hammad S, Sitek B, Roessler S, Ebert MP, et al. TKI-resistant human lung cancer cells in vitro and in vivo through inhibition Identification of the Consistently Altered Metabolic Targets in Human of IL-6 signaling and EMT reversal. Clin Cancer Res Off J Am Assoc Cancer Hepatocellular Carcinoma. Cell Mol Gastroenterol Hepatol. 2017;4:303–323.e1. Res. 2014;20:2714–26. 28. Shi H, Fang R, Li Y, Li L, Zhang W, Wang H, et al. The oncoprotein HBXIP 50. Blandino G, Valerio M, Cioce M, Mori F, Casadei L, Pulito C, et al. Metformin suppresses gluconeogenesis through modulating PCK1 to enhance the elicits anticancer effects through the sequential modulation of DICER and c- growth of hepatoma cells. Cancer Lett. 2016;382:147–56. MYC. Nat Commun. 2012;3:865. 29. Bian X-L, Chen H-Z, Yang P-B, Li Y-P, Zhang F-N, Zhang J-Y, et al. Nur77 51. Chen H-P, Shieh J-J, Chang C-C, Chen T-T, Lin J-T, Wu M-S, et al. Metformin suppresses hepatocellular carcinoma via switching glucose metabolism decreases hepatocellular carcinoma risk in a dose-dependent manner: toward gluconeogenesis through attenuating phosphoenolpyruvate population-based and in vitro studies. Gut. 2013;62:606–15. carboxykinase sumoylation. Nat Commun. 2017;8:14420. 30. Hawley SA, Davison M, Woods A, Davies SP, Beri RK, Carling D, et al. Characterization of the AMP-activated protein kinase kinase from rat liver and identification of threonine 172 as the major site at which it phosphorylates AMP-activated protein kinase. J Biol Chem. 1996;271:27879–87. 31. Samovski D, Sun J, Pietka T, Gross RW, Eckel RH, Su X, et al. Regulation of AMPK activation by CD36 links fatty acid uptake to β-oxidation. Diabetes. 2015;64:353–9. 32. Vink EI, Smiley JR, Mohr I. Subversion of Host Responses to Energy Insufficiency by Us3 Supports Herpes Simplex Virus 1 Replication during Stress. J Virol. 2017 [cited 2018 Jul 26];91. Available from: https://www.ncbi. nlm.nih.gov/pmc/articles/PMC5487550/ 33. Koo S-H, Flechner L, Qi L, Zhang X, Screaton RA, Jeffries S, et al. The CREB coactivator TORC2 is a key regulator of fasting glucose metabolism. Nature. 2005;437:1109–11. 34. Mihaylova MM, Vasquez DS, Ravnskjaer K, Denechaud P-D, Yu RT, Alvarez JG, et al. Class IIa histone deacetylases are hormone-activated regulators of FOXO and mammalian glucose homeostasis. Cell. 2011; 145:607–21. 35. Herzig S, Shaw RJ. AMPK: guardian of metabolism and mitochondrial homeostasis. Nat Rev Mol Cell Biol. 2018;19:121–35. 36. Zhang C-S, Hawley SA, Zong Y, Li M, Wang Z, Gray A, et al. Fructose-1,6- bisphosphate and aldolase mediate glucose sensing by AMPK. Nature. 2017; 548:112–6. 37. Xiao B, Sanders MJ, Carmena D, Bright NJ, Haire LF, Underwood E, et al. Structural basis of AMPK regulation by small molecule activators. Nat Commun. 2013;4:3017. 38. Xiao B, Sanders MJ, Underwood E, Heath R, Mayer FV, Carmena D, et al. Structure of mammalian AMPK and its regulation by ADP. Nature. 2011;472:230–3. 39. Woods A, Johnstone SR, Dickerson K, Leiper FC, Fryer LGD, Neumann D, et al. LKB1 is the upstream kinase in the AMP-activated protein kinase cascade. Curr Biol CB. 2003;13:2004–8. 40. Hawley SA, Pan DA, Mustard KJ, Ross L, Bain J, Edelman AM, et al. Calmodulin-dependent protein kinase kinase-beta is an alternative upstream kinase for AMP-activated protein kinase. Cell Metab. 2005;2:9–19. 41. Stark R, Pasquel F, Turcu A, Pongratz RL, Roden M, Cline GW, et al. Phosphoenolpyruvate cycling via mitochondrial phosphoenolpyruvate carboxykinase links anaplerosis and mitochondrial GTP with insulin secretion. J Biol Chem. 2009;284:26578–90. 42. Sy J. Reversibility of the Pyrophosphoryl transfer from ATP to GTP by Escherichia coli stringent factor. Proc Natl Acad Sci U S A. 1974;71:3470–3. 43. Muimo R, Crawford RM, Mehta A. Nucleoside diphosphate kinase a as a controller of AMP-kinase in airway epithelia. J Bioenerg Biomembr. 2006;38:181. 44. Rattan R, Giri S, Singh AK, Singh I. 5-Aminoimidazole-4-carboxamide-1-beta- D-ribofuranoside inhibits cancer cell proliferation in vitro and in vivo via AMP-activated protein kinase. J Biol Chem. 2005;280:39582–93. 45. Short JD, Dere R, Houston KD, Cai S-L, Kim J, Bergeron JM, et al. AMPK- mediated phosphorylation of murine p27 at T197 promotes binding of 14- 3-3 proteins and increases p27 stability. Mol Carcinog. 2010;49:429–39.